1*4882a593Smuzhiyun /* Generate assembler source containing symbol information
2*4882a593Smuzhiyun *
3*4882a593Smuzhiyun * Copyright 2002 by Kai Germaschewski
4*4882a593Smuzhiyun *
5*4882a593Smuzhiyun * This software may be used and distributed according to the terms
6*4882a593Smuzhiyun * of the GNU General Public License, incorporated herein by reference.
7*4882a593Smuzhiyun *
8*4882a593Smuzhiyun * Usage: nm -n vmlinux | scripts/kallsyms [--all-symbols] > symbols.S
9*4882a593Smuzhiyun *
10*4882a593Smuzhiyun * Table compression uses all the unused char codes on the symbols and
11*4882a593Smuzhiyun * maps these to the most used substrings (tokens). For instance, it might
12*4882a593Smuzhiyun * map char code 0xF7 to represent "write_" and then in every symbol where
13*4882a593Smuzhiyun * "write_" appears it can be replaced by 0xF7, saving 5 bytes.
14*4882a593Smuzhiyun * The used codes themselves are also placed in the table so that the
15*4882a593Smuzhiyun * decompresion can work without "special cases".
16*4882a593Smuzhiyun * Applied to kernel symbols, this usually produces a compression ratio
17*4882a593Smuzhiyun * of about 50%.
18*4882a593Smuzhiyun *
19*4882a593Smuzhiyun */
20*4882a593Smuzhiyun
21*4882a593Smuzhiyun #include <stdbool.h>
22*4882a593Smuzhiyun #include <stdio.h>
23*4882a593Smuzhiyun #include <stdlib.h>
24*4882a593Smuzhiyun #include <string.h>
25*4882a593Smuzhiyun #include <ctype.h>
26*4882a593Smuzhiyun #include <limits.h>
27*4882a593Smuzhiyun
28*4882a593Smuzhiyun #define ARRAY_SIZE(arr) (sizeof(arr) / sizeof(arr[0]))
29*4882a593Smuzhiyun
30*4882a593Smuzhiyun #define KSYM_NAME_LEN 128
31*4882a593Smuzhiyun
32*4882a593Smuzhiyun struct sym_entry {
33*4882a593Smuzhiyun unsigned long long addr;
34*4882a593Smuzhiyun unsigned int len;
35*4882a593Smuzhiyun unsigned int start_pos;
36*4882a593Smuzhiyun unsigned int percpu_absolute;
37*4882a593Smuzhiyun unsigned char sym[];
38*4882a593Smuzhiyun };
39*4882a593Smuzhiyun
40*4882a593Smuzhiyun struct addr_range {
41*4882a593Smuzhiyun const char *start_sym, *end_sym;
42*4882a593Smuzhiyun unsigned long long start, end;
43*4882a593Smuzhiyun };
44*4882a593Smuzhiyun
45*4882a593Smuzhiyun static unsigned long long _text;
46*4882a593Smuzhiyun static unsigned long long relative_base;
47*4882a593Smuzhiyun static struct addr_range text_ranges[] = {
48*4882a593Smuzhiyun { "_stext", "_etext" },
49*4882a593Smuzhiyun { "_sinittext", "_einittext" },
50*4882a593Smuzhiyun };
51*4882a593Smuzhiyun #define text_range_text (&text_ranges[0])
52*4882a593Smuzhiyun #define text_range_inittext (&text_ranges[1])
53*4882a593Smuzhiyun
54*4882a593Smuzhiyun static struct addr_range percpu_range = {
55*4882a593Smuzhiyun "__per_cpu_start", "__per_cpu_end", -1ULL, 0
56*4882a593Smuzhiyun };
57*4882a593Smuzhiyun
58*4882a593Smuzhiyun static struct sym_entry **table;
59*4882a593Smuzhiyun static unsigned int table_size, table_cnt;
60*4882a593Smuzhiyun static int all_symbols;
61*4882a593Smuzhiyun static int absolute_percpu;
62*4882a593Smuzhiyun static int base_relative;
63*4882a593Smuzhiyun
64*4882a593Smuzhiyun static int token_profit[0x10000];
65*4882a593Smuzhiyun
66*4882a593Smuzhiyun /* the table that holds the result of the compression */
67*4882a593Smuzhiyun static unsigned char best_table[256][2];
68*4882a593Smuzhiyun static unsigned char best_table_len[256];
69*4882a593Smuzhiyun
70*4882a593Smuzhiyun
usage(void)71*4882a593Smuzhiyun static void usage(void)
72*4882a593Smuzhiyun {
73*4882a593Smuzhiyun fprintf(stderr, "Usage: kallsyms [--all-symbols] "
74*4882a593Smuzhiyun "[--base-relative] < in.map > out.S\n");
75*4882a593Smuzhiyun exit(1);
76*4882a593Smuzhiyun }
77*4882a593Smuzhiyun
sym_name(const struct sym_entry * s)78*4882a593Smuzhiyun static char *sym_name(const struct sym_entry *s)
79*4882a593Smuzhiyun {
80*4882a593Smuzhiyun return (char *)s->sym + 1;
81*4882a593Smuzhiyun }
82*4882a593Smuzhiyun
is_ignored_symbol(const char * name,char type)83*4882a593Smuzhiyun static bool is_ignored_symbol(const char *name, char type)
84*4882a593Smuzhiyun {
85*4882a593Smuzhiyun /* Symbol names that exactly match to the following are ignored.*/
86*4882a593Smuzhiyun static const char * const ignored_symbols[] = {
87*4882a593Smuzhiyun /*
88*4882a593Smuzhiyun * Symbols which vary between passes. Passes 1 and 2 must have
89*4882a593Smuzhiyun * identical symbol lists. The kallsyms_* symbols below are
90*4882a593Smuzhiyun * only added after pass 1, they would be included in pass 2
91*4882a593Smuzhiyun * when --all-symbols is specified so exclude them to get a
92*4882a593Smuzhiyun * stable symbol list.
93*4882a593Smuzhiyun */
94*4882a593Smuzhiyun "kallsyms_addresses",
95*4882a593Smuzhiyun "kallsyms_offsets",
96*4882a593Smuzhiyun "kallsyms_relative_base",
97*4882a593Smuzhiyun "kallsyms_num_syms",
98*4882a593Smuzhiyun "kallsyms_names",
99*4882a593Smuzhiyun "kallsyms_markers",
100*4882a593Smuzhiyun "kallsyms_token_table",
101*4882a593Smuzhiyun "kallsyms_token_index",
102*4882a593Smuzhiyun /* Exclude linker generated symbols which vary between passes */
103*4882a593Smuzhiyun "_SDA_BASE_", /* ppc */
104*4882a593Smuzhiyun "_SDA2_BASE_", /* ppc */
105*4882a593Smuzhiyun NULL
106*4882a593Smuzhiyun };
107*4882a593Smuzhiyun
108*4882a593Smuzhiyun /* Symbol names that begin with the following are ignored.*/
109*4882a593Smuzhiyun static const char * const ignored_prefixes[] = {
110*4882a593Smuzhiyun "$", /* local symbols for ARM, MIPS, etc. */
111*4882a593Smuzhiyun ".LASANPC", /* s390 kasan local symbols */
112*4882a593Smuzhiyun "__crc_", /* modversions */
113*4882a593Smuzhiyun "__efistub_", /* arm64 EFI stub namespace */
114*4882a593Smuzhiyun "__kvm_nvhe_", /* arm64 non-VHE KVM namespace */
115*4882a593Smuzhiyun "__AArch64ADRPThunk_", /* arm64 lld */
116*4882a593Smuzhiyun "__ARMV5PILongThunk_", /* arm lld */
117*4882a593Smuzhiyun "__ARMV7PILongThunk_",
118*4882a593Smuzhiyun "__ThumbV7PILongThunk_",
119*4882a593Smuzhiyun "__LA25Thunk_", /* mips lld */
120*4882a593Smuzhiyun "__microLA25Thunk_",
121*4882a593Smuzhiyun NULL
122*4882a593Smuzhiyun };
123*4882a593Smuzhiyun
124*4882a593Smuzhiyun /* Symbol names that end with the following are ignored.*/
125*4882a593Smuzhiyun static const char * const ignored_suffixes[] = {
126*4882a593Smuzhiyun "_from_arm", /* arm */
127*4882a593Smuzhiyun "_from_thumb", /* arm */
128*4882a593Smuzhiyun "_veneer", /* arm */
129*4882a593Smuzhiyun NULL
130*4882a593Smuzhiyun };
131*4882a593Smuzhiyun
132*4882a593Smuzhiyun /* Symbol names that contain the following are ignored.*/
133*4882a593Smuzhiyun static const char * const ignored_matches[] = {
134*4882a593Smuzhiyun ".long_branch.", /* ppc stub */
135*4882a593Smuzhiyun ".plt_branch.", /* ppc stub */
136*4882a593Smuzhiyun NULL
137*4882a593Smuzhiyun };
138*4882a593Smuzhiyun
139*4882a593Smuzhiyun const char * const *p;
140*4882a593Smuzhiyun
141*4882a593Smuzhiyun for (p = ignored_symbols; *p; p++)
142*4882a593Smuzhiyun if (!strcmp(name, *p))
143*4882a593Smuzhiyun return true;
144*4882a593Smuzhiyun
145*4882a593Smuzhiyun for (p = ignored_prefixes; *p; p++)
146*4882a593Smuzhiyun if (!strncmp(name, *p, strlen(*p)))
147*4882a593Smuzhiyun return true;
148*4882a593Smuzhiyun
149*4882a593Smuzhiyun for (p = ignored_suffixes; *p; p++) {
150*4882a593Smuzhiyun int l = strlen(name) - strlen(*p);
151*4882a593Smuzhiyun
152*4882a593Smuzhiyun if (l >= 0 && !strcmp(name + l, *p))
153*4882a593Smuzhiyun return true;
154*4882a593Smuzhiyun }
155*4882a593Smuzhiyun
156*4882a593Smuzhiyun for (p = ignored_matches; *p; p++) {
157*4882a593Smuzhiyun if (strstr(name, *p))
158*4882a593Smuzhiyun return true;
159*4882a593Smuzhiyun }
160*4882a593Smuzhiyun
161*4882a593Smuzhiyun if (type == 'U' || type == 'u')
162*4882a593Smuzhiyun return true;
163*4882a593Smuzhiyun /* exclude debugging symbols */
164*4882a593Smuzhiyun if (type == 'N' || type == 'n')
165*4882a593Smuzhiyun return true;
166*4882a593Smuzhiyun
167*4882a593Smuzhiyun if (toupper(type) == 'A') {
168*4882a593Smuzhiyun /* Keep these useful absolute symbols */
169*4882a593Smuzhiyun if (strcmp(name, "__kernel_syscall_via_break") &&
170*4882a593Smuzhiyun strcmp(name, "__kernel_syscall_via_epc") &&
171*4882a593Smuzhiyun strcmp(name, "__kernel_sigtramp") &&
172*4882a593Smuzhiyun strcmp(name, "__gp"))
173*4882a593Smuzhiyun return true;
174*4882a593Smuzhiyun }
175*4882a593Smuzhiyun
176*4882a593Smuzhiyun return false;
177*4882a593Smuzhiyun }
178*4882a593Smuzhiyun
check_symbol_range(const char * sym,unsigned long long addr,struct addr_range * ranges,int entries)179*4882a593Smuzhiyun static void check_symbol_range(const char *sym, unsigned long long addr,
180*4882a593Smuzhiyun struct addr_range *ranges, int entries)
181*4882a593Smuzhiyun {
182*4882a593Smuzhiyun size_t i;
183*4882a593Smuzhiyun struct addr_range *ar;
184*4882a593Smuzhiyun
185*4882a593Smuzhiyun for (i = 0; i < entries; ++i) {
186*4882a593Smuzhiyun ar = &ranges[i];
187*4882a593Smuzhiyun
188*4882a593Smuzhiyun if (strcmp(sym, ar->start_sym) == 0) {
189*4882a593Smuzhiyun ar->start = addr;
190*4882a593Smuzhiyun return;
191*4882a593Smuzhiyun } else if (strcmp(sym, ar->end_sym) == 0) {
192*4882a593Smuzhiyun ar->end = addr;
193*4882a593Smuzhiyun return;
194*4882a593Smuzhiyun }
195*4882a593Smuzhiyun }
196*4882a593Smuzhiyun }
197*4882a593Smuzhiyun
read_symbol(FILE * in)198*4882a593Smuzhiyun static struct sym_entry *read_symbol(FILE *in)
199*4882a593Smuzhiyun {
200*4882a593Smuzhiyun char name[500], type;
201*4882a593Smuzhiyun unsigned long long addr;
202*4882a593Smuzhiyun unsigned int len;
203*4882a593Smuzhiyun struct sym_entry *sym;
204*4882a593Smuzhiyun int rc;
205*4882a593Smuzhiyun
206*4882a593Smuzhiyun rc = fscanf(in, "%llx %c %499s\n", &addr, &type, name);
207*4882a593Smuzhiyun if (rc != 3) {
208*4882a593Smuzhiyun if (rc != EOF && fgets(name, 500, in) == NULL)
209*4882a593Smuzhiyun fprintf(stderr, "Read error or end of file.\n");
210*4882a593Smuzhiyun return NULL;
211*4882a593Smuzhiyun }
212*4882a593Smuzhiyun if (strlen(name) >= KSYM_NAME_LEN) {
213*4882a593Smuzhiyun fprintf(stderr, "Symbol %s too long for kallsyms (%zu >= %d).\n"
214*4882a593Smuzhiyun "Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n",
215*4882a593Smuzhiyun name, strlen(name), KSYM_NAME_LEN);
216*4882a593Smuzhiyun return NULL;
217*4882a593Smuzhiyun }
218*4882a593Smuzhiyun
219*4882a593Smuzhiyun if (strcmp(name, "_text") == 0)
220*4882a593Smuzhiyun _text = addr;
221*4882a593Smuzhiyun
222*4882a593Smuzhiyun /* Ignore most absolute/undefined (?) symbols. */
223*4882a593Smuzhiyun if (is_ignored_symbol(name, type))
224*4882a593Smuzhiyun return NULL;
225*4882a593Smuzhiyun
226*4882a593Smuzhiyun check_symbol_range(name, addr, text_ranges, ARRAY_SIZE(text_ranges));
227*4882a593Smuzhiyun check_symbol_range(name, addr, &percpu_range, 1);
228*4882a593Smuzhiyun
229*4882a593Smuzhiyun /* include the type field in the symbol name, so that it gets
230*4882a593Smuzhiyun * compressed together */
231*4882a593Smuzhiyun
232*4882a593Smuzhiyun len = strlen(name) + 1;
233*4882a593Smuzhiyun
234*4882a593Smuzhiyun sym = malloc(sizeof(*sym) + len + 1);
235*4882a593Smuzhiyun if (!sym) {
236*4882a593Smuzhiyun fprintf(stderr, "kallsyms failure: "
237*4882a593Smuzhiyun "unable to allocate required amount of memory\n");
238*4882a593Smuzhiyun exit(EXIT_FAILURE);
239*4882a593Smuzhiyun }
240*4882a593Smuzhiyun sym->addr = addr;
241*4882a593Smuzhiyun sym->len = len;
242*4882a593Smuzhiyun sym->sym[0] = type;
243*4882a593Smuzhiyun strcpy(sym_name(sym), name);
244*4882a593Smuzhiyun sym->percpu_absolute = 0;
245*4882a593Smuzhiyun
246*4882a593Smuzhiyun return sym;
247*4882a593Smuzhiyun }
248*4882a593Smuzhiyun
symbol_in_range(const struct sym_entry * s,const struct addr_range * ranges,int entries)249*4882a593Smuzhiyun static int symbol_in_range(const struct sym_entry *s,
250*4882a593Smuzhiyun const struct addr_range *ranges, int entries)
251*4882a593Smuzhiyun {
252*4882a593Smuzhiyun size_t i;
253*4882a593Smuzhiyun const struct addr_range *ar;
254*4882a593Smuzhiyun
255*4882a593Smuzhiyun for (i = 0; i < entries; ++i) {
256*4882a593Smuzhiyun ar = &ranges[i];
257*4882a593Smuzhiyun
258*4882a593Smuzhiyun if (s->addr >= ar->start && s->addr <= ar->end)
259*4882a593Smuzhiyun return 1;
260*4882a593Smuzhiyun }
261*4882a593Smuzhiyun
262*4882a593Smuzhiyun return 0;
263*4882a593Smuzhiyun }
264*4882a593Smuzhiyun
symbol_valid(const struct sym_entry * s)265*4882a593Smuzhiyun static int symbol_valid(const struct sym_entry *s)
266*4882a593Smuzhiyun {
267*4882a593Smuzhiyun const char *name = sym_name(s);
268*4882a593Smuzhiyun
269*4882a593Smuzhiyun /* if --all-symbols is not specified, then symbols outside the text
270*4882a593Smuzhiyun * and inittext sections are discarded */
271*4882a593Smuzhiyun if (!all_symbols) {
272*4882a593Smuzhiyun if (symbol_in_range(s, text_ranges,
273*4882a593Smuzhiyun ARRAY_SIZE(text_ranges)) == 0)
274*4882a593Smuzhiyun return 0;
275*4882a593Smuzhiyun /* Corner case. Discard any symbols with the same value as
276*4882a593Smuzhiyun * _etext _einittext; they can move between pass 1 and 2 when
277*4882a593Smuzhiyun * the kallsyms data are added. If these symbols move then
278*4882a593Smuzhiyun * they may get dropped in pass 2, which breaks the kallsyms
279*4882a593Smuzhiyun * rules.
280*4882a593Smuzhiyun */
281*4882a593Smuzhiyun if ((s->addr == text_range_text->end &&
282*4882a593Smuzhiyun strcmp(name, text_range_text->end_sym)) ||
283*4882a593Smuzhiyun (s->addr == text_range_inittext->end &&
284*4882a593Smuzhiyun strcmp(name, text_range_inittext->end_sym)))
285*4882a593Smuzhiyun return 0;
286*4882a593Smuzhiyun }
287*4882a593Smuzhiyun
288*4882a593Smuzhiyun return 1;
289*4882a593Smuzhiyun }
290*4882a593Smuzhiyun
291*4882a593Smuzhiyun /* remove all the invalid symbols from the table */
shrink_table(void)292*4882a593Smuzhiyun static void shrink_table(void)
293*4882a593Smuzhiyun {
294*4882a593Smuzhiyun unsigned int i, pos;
295*4882a593Smuzhiyun
296*4882a593Smuzhiyun pos = 0;
297*4882a593Smuzhiyun for (i = 0; i < table_cnt; i++) {
298*4882a593Smuzhiyun if (symbol_valid(table[i])) {
299*4882a593Smuzhiyun if (pos != i)
300*4882a593Smuzhiyun table[pos] = table[i];
301*4882a593Smuzhiyun pos++;
302*4882a593Smuzhiyun } else {
303*4882a593Smuzhiyun free(table[i]);
304*4882a593Smuzhiyun }
305*4882a593Smuzhiyun }
306*4882a593Smuzhiyun table_cnt = pos;
307*4882a593Smuzhiyun
308*4882a593Smuzhiyun /* When valid symbol is not registered, exit to error */
309*4882a593Smuzhiyun if (!table_cnt) {
310*4882a593Smuzhiyun fprintf(stderr, "No valid symbol.\n");
311*4882a593Smuzhiyun exit(1);
312*4882a593Smuzhiyun }
313*4882a593Smuzhiyun }
314*4882a593Smuzhiyun
read_map(FILE * in)315*4882a593Smuzhiyun static void read_map(FILE *in)
316*4882a593Smuzhiyun {
317*4882a593Smuzhiyun struct sym_entry *sym;
318*4882a593Smuzhiyun
319*4882a593Smuzhiyun while (!feof(in)) {
320*4882a593Smuzhiyun sym = read_symbol(in);
321*4882a593Smuzhiyun if (!sym)
322*4882a593Smuzhiyun continue;
323*4882a593Smuzhiyun
324*4882a593Smuzhiyun sym->start_pos = table_cnt;
325*4882a593Smuzhiyun
326*4882a593Smuzhiyun if (table_cnt >= table_size) {
327*4882a593Smuzhiyun table_size += 10000;
328*4882a593Smuzhiyun table = realloc(table, sizeof(*table) * table_size);
329*4882a593Smuzhiyun if (!table) {
330*4882a593Smuzhiyun fprintf(stderr, "out of memory\n");
331*4882a593Smuzhiyun exit (1);
332*4882a593Smuzhiyun }
333*4882a593Smuzhiyun }
334*4882a593Smuzhiyun
335*4882a593Smuzhiyun table[table_cnt++] = sym;
336*4882a593Smuzhiyun }
337*4882a593Smuzhiyun }
338*4882a593Smuzhiyun
output_label(const char * label)339*4882a593Smuzhiyun static void output_label(const char *label)
340*4882a593Smuzhiyun {
341*4882a593Smuzhiyun printf(".globl %s\n", label);
342*4882a593Smuzhiyun printf("\tALGN\n");
343*4882a593Smuzhiyun printf("%s:\n", label);
344*4882a593Smuzhiyun }
345*4882a593Smuzhiyun
346*4882a593Smuzhiyun /* Provide proper symbols relocatability by their '_text' relativeness. */
output_address(unsigned long long addr)347*4882a593Smuzhiyun static void output_address(unsigned long long addr)
348*4882a593Smuzhiyun {
349*4882a593Smuzhiyun if (_text <= addr)
350*4882a593Smuzhiyun printf("\tPTR\t_text + %#llx\n", addr - _text);
351*4882a593Smuzhiyun else
352*4882a593Smuzhiyun printf("\tPTR\t_text - %#llx\n", _text - addr);
353*4882a593Smuzhiyun }
354*4882a593Smuzhiyun
355*4882a593Smuzhiyun /* uncompress a compressed symbol. When this function is called, the best table
356*4882a593Smuzhiyun * might still be compressed itself, so the function needs to be recursive */
expand_symbol(const unsigned char * data,int len,char * result)357*4882a593Smuzhiyun static int expand_symbol(const unsigned char *data, int len, char *result)
358*4882a593Smuzhiyun {
359*4882a593Smuzhiyun int c, rlen, total=0;
360*4882a593Smuzhiyun
361*4882a593Smuzhiyun while (len) {
362*4882a593Smuzhiyun c = *data;
363*4882a593Smuzhiyun /* if the table holds a single char that is the same as the one
364*4882a593Smuzhiyun * we are looking for, then end the search */
365*4882a593Smuzhiyun if (best_table[c][0]==c && best_table_len[c]==1) {
366*4882a593Smuzhiyun *result++ = c;
367*4882a593Smuzhiyun total++;
368*4882a593Smuzhiyun } else {
369*4882a593Smuzhiyun /* if not, recurse and expand */
370*4882a593Smuzhiyun rlen = expand_symbol(best_table[c], best_table_len[c], result);
371*4882a593Smuzhiyun total += rlen;
372*4882a593Smuzhiyun result += rlen;
373*4882a593Smuzhiyun }
374*4882a593Smuzhiyun data++;
375*4882a593Smuzhiyun len--;
376*4882a593Smuzhiyun }
377*4882a593Smuzhiyun *result=0;
378*4882a593Smuzhiyun
379*4882a593Smuzhiyun return total;
380*4882a593Smuzhiyun }
381*4882a593Smuzhiyun
symbol_absolute(const struct sym_entry * s)382*4882a593Smuzhiyun static int symbol_absolute(const struct sym_entry *s)
383*4882a593Smuzhiyun {
384*4882a593Smuzhiyun return s->percpu_absolute;
385*4882a593Smuzhiyun }
386*4882a593Smuzhiyun
write_src(void)387*4882a593Smuzhiyun static void write_src(void)
388*4882a593Smuzhiyun {
389*4882a593Smuzhiyun unsigned int i, k, off;
390*4882a593Smuzhiyun unsigned int best_idx[256];
391*4882a593Smuzhiyun unsigned int *markers;
392*4882a593Smuzhiyun char buf[KSYM_NAME_LEN];
393*4882a593Smuzhiyun
394*4882a593Smuzhiyun printf("#include <asm/bitsperlong.h>\n");
395*4882a593Smuzhiyun printf("#if BITS_PER_LONG == 64\n");
396*4882a593Smuzhiyun printf("#define PTR .quad\n");
397*4882a593Smuzhiyun printf("#define ALGN .balign 8\n");
398*4882a593Smuzhiyun printf("#else\n");
399*4882a593Smuzhiyun printf("#define PTR .long\n");
400*4882a593Smuzhiyun printf("#define ALGN .balign 4\n");
401*4882a593Smuzhiyun printf("#endif\n");
402*4882a593Smuzhiyun
403*4882a593Smuzhiyun printf("\t.section .rodata, \"a\"\n");
404*4882a593Smuzhiyun
405*4882a593Smuzhiyun if (!base_relative)
406*4882a593Smuzhiyun output_label("kallsyms_addresses");
407*4882a593Smuzhiyun else
408*4882a593Smuzhiyun output_label("kallsyms_offsets");
409*4882a593Smuzhiyun
410*4882a593Smuzhiyun for (i = 0; i < table_cnt; i++) {
411*4882a593Smuzhiyun if (base_relative) {
412*4882a593Smuzhiyun /*
413*4882a593Smuzhiyun * Use the offset relative to the lowest value
414*4882a593Smuzhiyun * encountered of all relative symbols, and emit
415*4882a593Smuzhiyun * non-relocatable fixed offsets that will be fixed
416*4882a593Smuzhiyun * up at runtime.
417*4882a593Smuzhiyun */
418*4882a593Smuzhiyun
419*4882a593Smuzhiyun long long offset;
420*4882a593Smuzhiyun int overflow;
421*4882a593Smuzhiyun
422*4882a593Smuzhiyun if (!absolute_percpu) {
423*4882a593Smuzhiyun offset = table[i]->addr - relative_base;
424*4882a593Smuzhiyun overflow = (offset < 0 || offset > UINT_MAX);
425*4882a593Smuzhiyun } else if (symbol_absolute(table[i])) {
426*4882a593Smuzhiyun offset = table[i]->addr;
427*4882a593Smuzhiyun overflow = (offset < 0 || offset > INT_MAX);
428*4882a593Smuzhiyun } else {
429*4882a593Smuzhiyun offset = relative_base - table[i]->addr - 1;
430*4882a593Smuzhiyun overflow = (offset < INT_MIN || offset >= 0);
431*4882a593Smuzhiyun }
432*4882a593Smuzhiyun if (overflow) {
433*4882a593Smuzhiyun fprintf(stderr, "kallsyms failure: "
434*4882a593Smuzhiyun "%s symbol value %#llx out of range in relative mode\n",
435*4882a593Smuzhiyun symbol_absolute(table[i]) ? "absolute" : "relative",
436*4882a593Smuzhiyun table[i]->addr);
437*4882a593Smuzhiyun exit(EXIT_FAILURE);
438*4882a593Smuzhiyun }
439*4882a593Smuzhiyun printf("\t.long\t%#x\n", (int)offset);
440*4882a593Smuzhiyun } else if (!symbol_absolute(table[i])) {
441*4882a593Smuzhiyun output_address(table[i]->addr);
442*4882a593Smuzhiyun } else {
443*4882a593Smuzhiyun printf("\tPTR\t%#llx\n", table[i]->addr);
444*4882a593Smuzhiyun }
445*4882a593Smuzhiyun }
446*4882a593Smuzhiyun printf("\n");
447*4882a593Smuzhiyun
448*4882a593Smuzhiyun if (base_relative) {
449*4882a593Smuzhiyun output_label("kallsyms_relative_base");
450*4882a593Smuzhiyun output_address(relative_base);
451*4882a593Smuzhiyun printf("\n");
452*4882a593Smuzhiyun }
453*4882a593Smuzhiyun
454*4882a593Smuzhiyun output_label("kallsyms_num_syms");
455*4882a593Smuzhiyun printf("\t.long\t%u\n", table_cnt);
456*4882a593Smuzhiyun printf("\n");
457*4882a593Smuzhiyun
458*4882a593Smuzhiyun /* table of offset markers, that give the offset in the compressed stream
459*4882a593Smuzhiyun * every 256 symbols */
460*4882a593Smuzhiyun markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
461*4882a593Smuzhiyun if (!markers) {
462*4882a593Smuzhiyun fprintf(stderr, "kallsyms failure: "
463*4882a593Smuzhiyun "unable to allocate required memory\n");
464*4882a593Smuzhiyun exit(EXIT_FAILURE);
465*4882a593Smuzhiyun }
466*4882a593Smuzhiyun
467*4882a593Smuzhiyun output_label("kallsyms_names");
468*4882a593Smuzhiyun off = 0;
469*4882a593Smuzhiyun for (i = 0; i < table_cnt; i++) {
470*4882a593Smuzhiyun if ((i & 0xFF) == 0)
471*4882a593Smuzhiyun markers[i >> 8] = off;
472*4882a593Smuzhiyun
473*4882a593Smuzhiyun printf("\t.byte 0x%02x", table[i]->len);
474*4882a593Smuzhiyun for (k = 0; k < table[i]->len; k++)
475*4882a593Smuzhiyun printf(", 0x%02x", table[i]->sym[k]);
476*4882a593Smuzhiyun printf("\n");
477*4882a593Smuzhiyun
478*4882a593Smuzhiyun off += table[i]->len + 1;
479*4882a593Smuzhiyun }
480*4882a593Smuzhiyun printf("\n");
481*4882a593Smuzhiyun
482*4882a593Smuzhiyun output_label("kallsyms_markers");
483*4882a593Smuzhiyun for (i = 0; i < ((table_cnt + 255) >> 8); i++)
484*4882a593Smuzhiyun printf("\t.long\t%u\n", markers[i]);
485*4882a593Smuzhiyun printf("\n");
486*4882a593Smuzhiyun
487*4882a593Smuzhiyun free(markers);
488*4882a593Smuzhiyun
489*4882a593Smuzhiyun output_label("kallsyms_token_table");
490*4882a593Smuzhiyun off = 0;
491*4882a593Smuzhiyun for (i = 0; i < 256; i++) {
492*4882a593Smuzhiyun best_idx[i] = off;
493*4882a593Smuzhiyun expand_symbol(best_table[i], best_table_len[i], buf);
494*4882a593Smuzhiyun printf("\t.asciz\t\"%s\"\n", buf);
495*4882a593Smuzhiyun off += strlen(buf) + 1;
496*4882a593Smuzhiyun }
497*4882a593Smuzhiyun printf("\n");
498*4882a593Smuzhiyun
499*4882a593Smuzhiyun output_label("kallsyms_token_index");
500*4882a593Smuzhiyun for (i = 0; i < 256; i++)
501*4882a593Smuzhiyun printf("\t.short\t%d\n", best_idx[i]);
502*4882a593Smuzhiyun printf("\n");
503*4882a593Smuzhiyun }
504*4882a593Smuzhiyun
505*4882a593Smuzhiyun
506*4882a593Smuzhiyun /* table lookup compression functions */
507*4882a593Smuzhiyun
508*4882a593Smuzhiyun /* count all the possible tokens in a symbol */
learn_symbol(const unsigned char * symbol,int len)509*4882a593Smuzhiyun static void learn_symbol(const unsigned char *symbol, int len)
510*4882a593Smuzhiyun {
511*4882a593Smuzhiyun int i;
512*4882a593Smuzhiyun
513*4882a593Smuzhiyun for (i = 0; i < len - 1; i++)
514*4882a593Smuzhiyun token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
515*4882a593Smuzhiyun }
516*4882a593Smuzhiyun
517*4882a593Smuzhiyun /* decrease the count for all the possible tokens in a symbol */
forget_symbol(const unsigned char * symbol,int len)518*4882a593Smuzhiyun static void forget_symbol(const unsigned char *symbol, int len)
519*4882a593Smuzhiyun {
520*4882a593Smuzhiyun int i;
521*4882a593Smuzhiyun
522*4882a593Smuzhiyun for (i = 0; i < len - 1; i++)
523*4882a593Smuzhiyun token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
524*4882a593Smuzhiyun }
525*4882a593Smuzhiyun
526*4882a593Smuzhiyun /* do the initial token count */
build_initial_tok_table(void)527*4882a593Smuzhiyun static void build_initial_tok_table(void)
528*4882a593Smuzhiyun {
529*4882a593Smuzhiyun unsigned int i;
530*4882a593Smuzhiyun
531*4882a593Smuzhiyun for (i = 0; i < table_cnt; i++)
532*4882a593Smuzhiyun learn_symbol(table[i]->sym, table[i]->len);
533*4882a593Smuzhiyun }
534*4882a593Smuzhiyun
find_token(unsigned char * str,int len,const unsigned char * token)535*4882a593Smuzhiyun static unsigned char *find_token(unsigned char *str, int len,
536*4882a593Smuzhiyun const unsigned char *token)
537*4882a593Smuzhiyun {
538*4882a593Smuzhiyun int i;
539*4882a593Smuzhiyun
540*4882a593Smuzhiyun for (i = 0; i < len - 1; i++) {
541*4882a593Smuzhiyun if (str[i] == token[0] && str[i+1] == token[1])
542*4882a593Smuzhiyun return &str[i];
543*4882a593Smuzhiyun }
544*4882a593Smuzhiyun return NULL;
545*4882a593Smuzhiyun }
546*4882a593Smuzhiyun
547*4882a593Smuzhiyun /* replace a given token in all the valid symbols. Use the sampled symbols
548*4882a593Smuzhiyun * to update the counts */
compress_symbols(const unsigned char * str,int idx)549*4882a593Smuzhiyun static void compress_symbols(const unsigned char *str, int idx)
550*4882a593Smuzhiyun {
551*4882a593Smuzhiyun unsigned int i, len, size;
552*4882a593Smuzhiyun unsigned char *p1, *p2;
553*4882a593Smuzhiyun
554*4882a593Smuzhiyun for (i = 0; i < table_cnt; i++) {
555*4882a593Smuzhiyun
556*4882a593Smuzhiyun len = table[i]->len;
557*4882a593Smuzhiyun p1 = table[i]->sym;
558*4882a593Smuzhiyun
559*4882a593Smuzhiyun /* find the token on the symbol */
560*4882a593Smuzhiyun p2 = find_token(p1, len, str);
561*4882a593Smuzhiyun if (!p2) continue;
562*4882a593Smuzhiyun
563*4882a593Smuzhiyun /* decrease the counts for this symbol's tokens */
564*4882a593Smuzhiyun forget_symbol(table[i]->sym, len);
565*4882a593Smuzhiyun
566*4882a593Smuzhiyun size = len;
567*4882a593Smuzhiyun
568*4882a593Smuzhiyun do {
569*4882a593Smuzhiyun *p2 = idx;
570*4882a593Smuzhiyun p2++;
571*4882a593Smuzhiyun size -= (p2 - p1);
572*4882a593Smuzhiyun memmove(p2, p2 + 1, size);
573*4882a593Smuzhiyun p1 = p2;
574*4882a593Smuzhiyun len--;
575*4882a593Smuzhiyun
576*4882a593Smuzhiyun if (size < 2) break;
577*4882a593Smuzhiyun
578*4882a593Smuzhiyun /* find the token on the symbol */
579*4882a593Smuzhiyun p2 = find_token(p1, size, str);
580*4882a593Smuzhiyun
581*4882a593Smuzhiyun } while (p2);
582*4882a593Smuzhiyun
583*4882a593Smuzhiyun table[i]->len = len;
584*4882a593Smuzhiyun
585*4882a593Smuzhiyun /* increase the counts for this symbol's new tokens */
586*4882a593Smuzhiyun learn_symbol(table[i]->sym, len);
587*4882a593Smuzhiyun }
588*4882a593Smuzhiyun }
589*4882a593Smuzhiyun
590*4882a593Smuzhiyun /* search the token with the maximum profit */
find_best_token(void)591*4882a593Smuzhiyun static int find_best_token(void)
592*4882a593Smuzhiyun {
593*4882a593Smuzhiyun int i, best, bestprofit;
594*4882a593Smuzhiyun
595*4882a593Smuzhiyun bestprofit=-10000;
596*4882a593Smuzhiyun best = 0;
597*4882a593Smuzhiyun
598*4882a593Smuzhiyun for (i = 0; i < 0x10000; i++) {
599*4882a593Smuzhiyun if (token_profit[i] > bestprofit) {
600*4882a593Smuzhiyun best = i;
601*4882a593Smuzhiyun bestprofit = token_profit[i];
602*4882a593Smuzhiyun }
603*4882a593Smuzhiyun }
604*4882a593Smuzhiyun return best;
605*4882a593Smuzhiyun }
606*4882a593Smuzhiyun
607*4882a593Smuzhiyun /* this is the core of the algorithm: calculate the "best" table */
optimize_result(void)608*4882a593Smuzhiyun static void optimize_result(void)
609*4882a593Smuzhiyun {
610*4882a593Smuzhiyun int i, best;
611*4882a593Smuzhiyun
612*4882a593Smuzhiyun /* using the '\0' symbol last allows compress_symbols to use standard
613*4882a593Smuzhiyun * fast string functions */
614*4882a593Smuzhiyun for (i = 255; i >= 0; i--) {
615*4882a593Smuzhiyun
616*4882a593Smuzhiyun /* if this table slot is empty (it is not used by an actual
617*4882a593Smuzhiyun * original char code */
618*4882a593Smuzhiyun if (!best_table_len[i]) {
619*4882a593Smuzhiyun
620*4882a593Smuzhiyun /* find the token with the best profit value */
621*4882a593Smuzhiyun best = find_best_token();
622*4882a593Smuzhiyun if (token_profit[best] == 0)
623*4882a593Smuzhiyun break;
624*4882a593Smuzhiyun
625*4882a593Smuzhiyun /* place it in the "best" table */
626*4882a593Smuzhiyun best_table_len[i] = 2;
627*4882a593Smuzhiyun best_table[i][0] = best & 0xFF;
628*4882a593Smuzhiyun best_table[i][1] = (best >> 8) & 0xFF;
629*4882a593Smuzhiyun
630*4882a593Smuzhiyun /* replace this token in all the valid symbols */
631*4882a593Smuzhiyun compress_symbols(best_table[i], i);
632*4882a593Smuzhiyun }
633*4882a593Smuzhiyun }
634*4882a593Smuzhiyun }
635*4882a593Smuzhiyun
636*4882a593Smuzhiyun /* start by placing the symbols that are actually used on the table */
insert_real_symbols_in_table(void)637*4882a593Smuzhiyun static void insert_real_symbols_in_table(void)
638*4882a593Smuzhiyun {
639*4882a593Smuzhiyun unsigned int i, j, c;
640*4882a593Smuzhiyun
641*4882a593Smuzhiyun for (i = 0; i < table_cnt; i++) {
642*4882a593Smuzhiyun for (j = 0; j < table[i]->len; j++) {
643*4882a593Smuzhiyun c = table[i]->sym[j];
644*4882a593Smuzhiyun best_table[c][0]=c;
645*4882a593Smuzhiyun best_table_len[c]=1;
646*4882a593Smuzhiyun }
647*4882a593Smuzhiyun }
648*4882a593Smuzhiyun }
649*4882a593Smuzhiyun
optimize_token_table(void)650*4882a593Smuzhiyun static void optimize_token_table(void)
651*4882a593Smuzhiyun {
652*4882a593Smuzhiyun build_initial_tok_table();
653*4882a593Smuzhiyun
654*4882a593Smuzhiyun insert_real_symbols_in_table();
655*4882a593Smuzhiyun
656*4882a593Smuzhiyun optimize_result();
657*4882a593Smuzhiyun }
658*4882a593Smuzhiyun
659*4882a593Smuzhiyun /* guess for "linker script provide" symbol */
may_be_linker_script_provide_symbol(const struct sym_entry * se)660*4882a593Smuzhiyun static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
661*4882a593Smuzhiyun {
662*4882a593Smuzhiyun const char *symbol = sym_name(se);
663*4882a593Smuzhiyun int len = se->len - 1;
664*4882a593Smuzhiyun
665*4882a593Smuzhiyun if (len < 8)
666*4882a593Smuzhiyun return 0;
667*4882a593Smuzhiyun
668*4882a593Smuzhiyun if (symbol[0] != '_' || symbol[1] != '_')
669*4882a593Smuzhiyun return 0;
670*4882a593Smuzhiyun
671*4882a593Smuzhiyun /* __start_XXXXX */
672*4882a593Smuzhiyun if (!memcmp(symbol + 2, "start_", 6))
673*4882a593Smuzhiyun return 1;
674*4882a593Smuzhiyun
675*4882a593Smuzhiyun /* __stop_XXXXX */
676*4882a593Smuzhiyun if (!memcmp(symbol + 2, "stop_", 5))
677*4882a593Smuzhiyun return 1;
678*4882a593Smuzhiyun
679*4882a593Smuzhiyun /* __end_XXXXX */
680*4882a593Smuzhiyun if (!memcmp(symbol + 2, "end_", 4))
681*4882a593Smuzhiyun return 1;
682*4882a593Smuzhiyun
683*4882a593Smuzhiyun /* __XXXXX_start */
684*4882a593Smuzhiyun if (!memcmp(symbol + len - 6, "_start", 6))
685*4882a593Smuzhiyun return 1;
686*4882a593Smuzhiyun
687*4882a593Smuzhiyun /* __XXXXX_end */
688*4882a593Smuzhiyun if (!memcmp(symbol + len - 4, "_end", 4))
689*4882a593Smuzhiyun return 1;
690*4882a593Smuzhiyun
691*4882a593Smuzhiyun return 0;
692*4882a593Smuzhiyun }
693*4882a593Smuzhiyun
compare_symbols(const void * a,const void * b)694*4882a593Smuzhiyun static int compare_symbols(const void *a, const void *b)
695*4882a593Smuzhiyun {
696*4882a593Smuzhiyun const struct sym_entry *sa = *(const struct sym_entry **)a;
697*4882a593Smuzhiyun const struct sym_entry *sb = *(const struct sym_entry **)b;
698*4882a593Smuzhiyun int wa, wb;
699*4882a593Smuzhiyun
700*4882a593Smuzhiyun /* sort by address first */
701*4882a593Smuzhiyun if (sa->addr > sb->addr)
702*4882a593Smuzhiyun return 1;
703*4882a593Smuzhiyun if (sa->addr < sb->addr)
704*4882a593Smuzhiyun return -1;
705*4882a593Smuzhiyun
706*4882a593Smuzhiyun /* sort by "weakness" type */
707*4882a593Smuzhiyun wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
708*4882a593Smuzhiyun wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
709*4882a593Smuzhiyun if (wa != wb)
710*4882a593Smuzhiyun return wa - wb;
711*4882a593Smuzhiyun
712*4882a593Smuzhiyun /* sort by "linker script provide" type */
713*4882a593Smuzhiyun wa = may_be_linker_script_provide_symbol(sa);
714*4882a593Smuzhiyun wb = may_be_linker_script_provide_symbol(sb);
715*4882a593Smuzhiyun if (wa != wb)
716*4882a593Smuzhiyun return wa - wb;
717*4882a593Smuzhiyun
718*4882a593Smuzhiyun /* sort by the number of prefix underscores */
719*4882a593Smuzhiyun wa = strspn(sym_name(sa), "_");
720*4882a593Smuzhiyun wb = strspn(sym_name(sb), "_");
721*4882a593Smuzhiyun if (wa != wb)
722*4882a593Smuzhiyun return wa - wb;
723*4882a593Smuzhiyun
724*4882a593Smuzhiyun /* sort by initial order, so that other symbols are left undisturbed */
725*4882a593Smuzhiyun return sa->start_pos - sb->start_pos;
726*4882a593Smuzhiyun }
727*4882a593Smuzhiyun
sort_symbols(void)728*4882a593Smuzhiyun static void sort_symbols(void)
729*4882a593Smuzhiyun {
730*4882a593Smuzhiyun qsort(table, table_cnt, sizeof(table[0]), compare_symbols);
731*4882a593Smuzhiyun }
732*4882a593Smuzhiyun
make_percpus_absolute(void)733*4882a593Smuzhiyun static void make_percpus_absolute(void)
734*4882a593Smuzhiyun {
735*4882a593Smuzhiyun unsigned int i;
736*4882a593Smuzhiyun
737*4882a593Smuzhiyun for (i = 0; i < table_cnt; i++)
738*4882a593Smuzhiyun if (symbol_in_range(table[i], &percpu_range, 1)) {
739*4882a593Smuzhiyun /*
740*4882a593Smuzhiyun * Keep the 'A' override for percpu symbols to
741*4882a593Smuzhiyun * ensure consistent behavior compared to older
742*4882a593Smuzhiyun * versions of this tool.
743*4882a593Smuzhiyun */
744*4882a593Smuzhiyun table[i]->sym[0] = 'A';
745*4882a593Smuzhiyun table[i]->percpu_absolute = 1;
746*4882a593Smuzhiyun }
747*4882a593Smuzhiyun }
748*4882a593Smuzhiyun
749*4882a593Smuzhiyun /* find the minimum non-absolute symbol address */
record_relative_base(void)750*4882a593Smuzhiyun static void record_relative_base(void)
751*4882a593Smuzhiyun {
752*4882a593Smuzhiyun unsigned int i;
753*4882a593Smuzhiyun
754*4882a593Smuzhiyun for (i = 0; i < table_cnt; i++)
755*4882a593Smuzhiyun if (!symbol_absolute(table[i])) {
756*4882a593Smuzhiyun /*
757*4882a593Smuzhiyun * The table is sorted by address.
758*4882a593Smuzhiyun * Take the first non-absolute symbol value.
759*4882a593Smuzhiyun */
760*4882a593Smuzhiyun relative_base = table[i]->addr;
761*4882a593Smuzhiyun return;
762*4882a593Smuzhiyun }
763*4882a593Smuzhiyun }
764*4882a593Smuzhiyun
main(int argc,char ** argv)765*4882a593Smuzhiyun int main(int argc, char **argv)
766*4882a593Smuzhiyun {
767*4882a593Smuzhiyun if (argc >= 2) {
768*4882a593Smuzhiyun int i;
769*4882a593Smuzhiyun for (i = 1; i < argc; i++) {
770*4882a593Smuzhiyun if(strcmp(argv[i], "--all-symbols") == 0)
771*4882a593Smuzhiyun all_symbols = 1;
772*4882a593Smuzhiyun else if (strcmp(argv[i], "--absolute-percpu") == 0)
773*4882a593Smuzhiyun absolute_percpu = 1;
774*4882a593Smuzhiyun else if (strcmp(argv[i], "--base-relative") == 0)
775*4882a593Smuzhiyun base_relative = 1;
776*4882a593Smuzhiyun else
777*4882a593Smuzhiyun usage();
778*4882a593Smuzhiyun }
779*4882a593Smuzhiyun } else if (argc != 1)
780*4882a593Smuzhiyun usage();
781*4882a593Smuzhiyun
782*4882a593Smuzhiyun read_map(stdin);
783*4882a593Smuzhiyun shrink_table();
784*4882a593Smuzhiyun if (absolute_percpu)
785*4882a593Smuzhiyun make_percpus_absolute();
786*4882a593Smuzhiyun sort_symbols();
787*4882a593Smuzhiyun if (base_relative)
788*4882a593Smuzhiyun record_relative_base();
789*4882a593Smuzhiyun optimize_token_table();
790*4882a593Smuzhiyun write_src();
791*4882a593Smuzhiyun
792*4882a593Smuzhiyun return 0;
793*4882a593Smuzhiyun }
794